Textractor_test/extensions/googletranslate.cpp

179 lines
4.6 KiB
C++
Raw Normal View History

2019-02-11 10:46:39 +08:00
#include "extension.h"
2018-12-19 05:32:28 +08:00
#include "text.h"
2019-02-11 10:46:39 +08:00
#include "util.h"
#include "network.h"
2018-11-04 11:26:53 +08:00
#include <ctime>
#include <QInputDialog>
#include <QTimer>
QStringList languages
{
"English: en",
"Afrikaans: af",
"Arabic: ar",
"Albanian: sq",
"Belarusian: be",
"Bengali: bn",
"Bosnian: bs",
"Bulgarian: bg",
"Catalan: ca",
"Chinese(Simplified): zh-CH",
"Chinese(Traditional): zh-TW",
"Croatian: hr",
"Czech: cs",
"Danish: da",
"Dutch: nl",
"Esperanto: eo",
"Estonian: et",
"Filipino: tl",
"Finnish: fi",
"French: fr",
"Galician: gl",
"German: de",
"Greek: el",
"Hebrew: iw",
"Hindi: hi",
"Hungarian: hu",
"Icelandic: is",
"Indonesian: id",
"Irish: ga",
"Italian: it",
"Japanese: ja",
"Klingon: tlh",
"Korean: ko",
"Latin: la",
"Latvian: lv",
"Lithuanian: lt",
"Macedonian: mk",
"Malay: ms",
"Maltese: mt",
"Norwegian: no",
"Persian: fa",
"Polish: pl",
"Portuguese: pt",
"Romanian: ro",
"Russian: ru",
"Serbian: sr",
"Slovak: sk",
"Slovenian: sl",
"Somali: so",
"Spanish: es",
"Swahili: sw",
"Swedish: sv",
"Thai: th",
"Turkish: tr",
"Ukranian: uk",
"Urdu: ur",
"Vietnamese: vi",
"Welsh: cy",
"Yiddish: yi",
"Zulu: zu"
};
2019-02-11 10:46:39 +08:00
std::wstring translateTo = L"en";
2018-11-04 11:26:53 +08:00
BOOL WINAPI DllMain(HMODULE hModule, DWORD ul_reason_for_call, LPVOID lpReserved)
{
switch (ul_reason_for_call)
{
case DLL_PROCESS_ATTACH:
{
QTimer::singleShot(0, []
{
2018-11-04 17:31:49 +08:00
translateTo = QInputDialog::getItem(
nullptr,
2018-12-19 05:32:28 +08:00
SELECT_LANGUAGE,
GOOGLE_PROMPT,
2018-11-04 17:31:49 +08:00
languages,
0, false, nullptr,
Qt::WindowCloseButtonHint
).split(" ")[1].toStdWString();
2018-11-04 11:26:53 +08:00
});
}
break;
case DLL_PROCESS_DETACH:
{
}
break;
}
return TRUE;
}
2019-02-11 10:46:39 +08:00
std::wstring GetTranslationUri(const std::wstring& text, unsigned TKK)
2018-11-04 11:26:53 +08:00
{
2018-11-08 18:51:20 +08:00
// If no TKK available, use this uri. Can't use too much or google will detect unauthorized access
2018-11-04 11:26:53 +08:00
if (!TKK) return L"/translate_a/single?client=gtx&dt=ld&dt=rm&dt=t&tl=" + translateTo + L"&q=" + text;
// Artikash 8/19/2018: reverse engineered from translate.google.com
2019-02-11 10:46:39 +08:00
std::wstring escapedText;
unsigned a = _time64(NULL) / 3600, b = a; // <- the first part of TKK
for (unsigned char ch : WideStringToString(text))
2018-11-04 11:26:53 +08:00
{
2019-02-11 10:46:39 +08:00
wchar_t escapedChar[4] = {};
swprintf_s<4>(escapedChar, L"%%%02X", (int)ch);
escapedText += escapedChar;
a += ch;
2018-11-04 11:26:53 +08:00
a += a << 10;
a ^= a >> 6;
}
a += a << 3;
a ^= a >> 11;
a += a << 15;
a ^= TKK;
a %= 1000000;
b ^= a;
2019-02-11 10:46:39 +08:00
return L"/translate_a/single?client=t&dt=ld&dt=rm&dt=t&tl=" + translateTo + L"&tk=" + std::to_wstring(a) + L"." + std::to_wstring(b) + L"&q=" + escapedText;
2018-11-04 11:26:53 +08:00
}
bool ProcessSentence(std::wstring& sentence, SentenceInfo sentenceInfo)
{
2019-02-18 08:14:49 +08:00
if (sentenceInfo["text number"] == 0) return false;
2018-11-04 11:26:53 +08:00
2019-01-24 02:41:50 +08:00
static std::atomic<HINTERNET> internet = NULL;
if (!internet) internet = WinHttpOpen(L"Mozilla/5.0 Textractor", WINHTTP_ACCESS_TYPE_DEFAULT_PROXY, NULL, NULL, 0);
2019-01-24 02:41:50 +08:00
static std::atomic<unsigned> TKK = 0;
2019-02-11 10:46:39 +08:00
static RateLimiter rateLimiter(30, 60 * 1000);
2018-11-04 11:26:53 +08:00
std::wstring translation;
2019-02-11 10:46:39 +08:00
if (!(rateLimiter.Request() || sentenceInfo["current select"])) translation = TOO_MANY_TRANS_REQUESTS;
else if (internet)
2018-11-04 11:26:53 +08:00
{
2018-12-19 03:21:27 +08:00
if (!TKK)
2019-02-11 10:46:39 +08:00
if (InternetHandle connection = WinHttpConnect(internet, L"translate.google.com", INTERNET_DEFAULT_HTTPS_PORT, 0))
if (InternetHandle request = WinHttpOpenRequest(connection, L"GET", L"/", NULL, NULL, NULL, WINHTTP_FLAG_SECURE))
2018-12-19 03:21:27 +08:00
if (WinHttpSendRequest(request, NULL, 0, NULL, 0, 0, NULL))
2019-02-11 10:46:39 +08:00
if (auto response = ReceiveHttpRequest(request))
if (std::wsmatch results; std::regex_search(response.value(), results, std::wregex(L"(\\d{7,})'"))) TKK = stoll(results[1]);
2018-11-04 11:26:53 +08:00
2019-02-11 10:46:39 +08:00
if (InternetHandle connection = WinHttpConnect(internet, L"translate.google.com", INTERNET_DEFAULT_HTTPS_PORT, 0))
if (InternetHandle request = WinHttpOpenRequest(connection, L"GET", GetTranslationUri(sentence, TKK).c_str(), NULL, NULL, NULL, WINHTTP_FLAG_ESCAPE_DISABLE | WINHTTP_FLAG_SECURE))
2018-11-04 11:26:53 +08:00
if (WinHttpSendRequest(request, NULL, 0, NULL, 0, 0, NULL))
2019-02-11 10:46:39 +08:00
if (auto response = ReceiveHttpRequest(request))
2019-02-17 11:43:31 +08:00
// Response formatted as JSON: starts with [[["
2019-02-11 10:46:39 +08:00
if (response.value()[0] == L'[')
2019-02-17 11:43:31 +08:00
{
2019-02-11 10:46:39 +08:00
for (std::wsmatch results; std::regex_search(response.value(), results, std::wregex(L"\\[\"(.*?)\",[n\"]")); response = results.suffix())
translation += std::wstring(results[1]) + L" ";
2019-02-17 11:43:31 +08:00
Escape(translation);
}
else
{
translation = TRANSLATION_ERROR + (L" (TKK=" + std::to_wstring(TKK) + L")");
TKK = 0;
}
2018-11-04 11:26:53 +08:00
}
2018-12-21 00:02:03 +08:00
if (translation.empty()) translation = TRANSLATION_ERROR;
2019-01-11 10:51:20 +08:00
sentence += L"\n" + translation;
2018-11-04 11:26:53 +08:00
return true;
2018-12-19 05:32:28 +08:00
}
2019-02-11 10:46:39 +08:00
TEST(
{
std::wstring test = L"こんにちは";
2019-02-13 12:54:33 +08:00
ProcessSentence(test, { SentenceInfo::DUMMY });
2019-02-11 10:46:39 +08:00
assert(test.find(L"Hello") != std::wstring::npos);
}
);